Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Mila Kashcheeva <kashcheeva.mila@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Re: xtpqml clustering |
Date | Mon, 7 Oct 2013 17:02:41 +0800 |
Dear Statalist, I work with a panel datatset where I regress firm-country imports on a number of parameters. I use xtpoisson function to be able to account for zero imports, use fixed effects and cluster standard errors. I want to include country, year and firm fixed effect and at the same time cluster standard errors by country. The xtpoisson command allows clustering only at a level of fixed effects included via "fe i(variable) vce(robust)" , and since I want to cluster by country I include ",fe i(id_country) vce(robust)" and also include year fixed effects and firm fixed effect manually in the regression. The regression calculates the estimates for all my main variables and for the year fixed effects, but does not calculate standard errors for the firm fixed effects and instead gives me dots for most of the firm fixed effects (with just a few standard errors calculated). Can someone explain why the xtpoisson command does not calculate standard errors for firm fixed effects and whether I can trust the rest of fixed effects in this regression. Thank you for you suggestion! Mila On Thu, Sep 5, 2013 at 1:22 PM, Mila Kashcheeva <kashcheeva.mila@gmail.com> wrote: > Dear Statalist, > > I work with a panel datatset where I regress firm-country imports on a > number of parameters. I use xtpqml function to be able to account for > zero imports, use fixed effects and cluster std. errors. I want to > cluster standard errors by country while also including year, country > and firm fixed effects. I do not understand two points. 1) why xtpqml > does not allow me to cluster by country when I include firm fixed > effects, it gives me the message, group variable i must be nested > within clusters; 2) when I cluster by country most of my country fe > are dropped, the message for the dropped fixed effects is, dropped > because it is constant within group. I wonder if it is a bug in the > program or a certain feature of xtpqml function which I do not > understand. > > Thank you, > Mila -- Best regards, Mila Kashcheeva Mila Kashcheeva, PhD Research Fellow http://www.ide.go.jp/English/Researchers/Kashcheeva_Mila_en.html Institute of Developing Economies, Japan External Trade Organization (IDE-JETRO) http://www.ide.go.jp/English/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/